Advancing Self-Consistency in ab initio DMET for Correlated Materials and Superconductivity
ORAL
Abstract
Ab initio Density Matrix Embedding Theory (DMET) has successfully reproduced superconducting trends observed experimentally in doped cuprates. However, how different self-consistency schemes affect predictions has not been systematically studied. Proper self-consistency is essential for capturing superconductivity beyond mean-field approaches.
We introduce several updates that improve the self-consistency of DMET, including a self-determined regularization parameter constrained by entropy. These developments make DMET more rigorous and robust. Benchmark results on both model and ab initio systems show that the improved scheme retains previously established trends while enhancing numerical stability and transferability. Ongoing work extends these applications to more complicated superconducting materials.
We introduce several updates that improve the self-consistency of DMET, including a self-determined regularization parameter constrained by entropy. These developments make DMET more rigorous and robust. Benchmark results on both model and ab initio systems show that the improved scheme retains previously established trends while enhancing numerical stability and transferability. Ongoing work extends these applications to more complicated superconducting materials.
*This work is primarily supported by the US Department of Energy, Office of Science, via grant no. DE SC0018140. G.K.-L.C. is a Simons Investigator in Physics. Calculations were performed using the facilities of National Energy Research Scientific Computing Center (NERSC), a U.S. Department of Energy Office of Science User Facility located at Lawrence Berkeley National Laboratory, under NERSC award ERCAP0023924, and in the Resnick High Performance Computing Center, supported by the Resnick Sustainability Institute at Caltech.
–
Presenters
-
Shunyue Yuan
- Caltech